AURORA – An AI-Augmented Code Refactoring Advisor

Abstract

Modern software systems often suffer from poor readability, high complexity, and increased technical debt due to inefficient coding practices. Manual code refactoring is time consuming and highly dependent on developer expertise, while existing automated tools are mostly rule-based and lack contextual understanding. To address these limitations, this paper presents AURORA, an AI-Augmented Code Refactoring Advisor that combines static program analysis with transformer-based artificial intelligence models to deliver safe and optimized refactoring suggestions. The proposed system employs Abstract Syntax Tree (AST) analysis to structurally understand source code and identify refactoring opportunities without altering program semantics. AI augmentation is applied to improve code quality by reducing redundancy and enhancing readability while preserving syntactic and logical correctness. A Python-based FastAPI backend performs code analysis and refactoring, while a Visual Studio Code extension provides real-time interaction for developers. Experimental evaluation on sample Python programs demonstrates effective reduction in code complexity and improved maintainability. AURORA offers a reliable and developer-friendly solution for intelligent code refactoring.